文献类型：期刊文献

中文题名：基于BP神经网络的弯管机回弹量预测

英文题名：Bending machine springback prediction based on BP neural network

作者：田娥[1];孙建东[1];刘自萍[1];李立新[1]

第一作者：田娥

机构：[1]北京联合大学机电学院

第一机构：北京联合大学机器人学院

年份：2016

卷号：0

期号：3

起止页码：70-73

中文期刊名：现代制造工程

外文期刊名：Modern Manufacturing Engineering

收录：CSTPCD;;北大核心:【北大核心2014】；CSCD:【CSCD_E2015_2016】；

基金：北京市教育委员会科技计划面上项目(SQKM201411417007)

语种：中文

中文关键词：大管径厚管壁管材;回弹量预测;模拟分析;BP神经网络

外文关键词：large-diameter and thick wall tube;springback prediction;simulation analysis;BP neural network

摘要：建筑钢结构不同管材的弯曲曲率变化较大,施工时需制作大量胎架,即费时费料又占用场地,故设计了一种冷弯机结构,基于该冷弯机,利用Midas7.8软件对建筑钢结构用管材进行冷弯成形模拟,基于神经网络建立3个输入参数的回弹量数据模型,选择1 000组模拟数据作为训练数据训练神经网络,500组模拟数据作为测试数据测试网络,将预测结果和样本结果进行比较和分析,结果表明,所建立的神经网络预测模型满足误差要求,可以用来预测大管径厚管壁管材冷弯成形后的回弹量。该项研究为开发具有自适应回弹量补偿性能的数控弯管系统提供理论基础。
Construction steel tubes curvatures differ greatly, when constructing, need to make a lot of tire racks, time and material consuming and space occupied. Design a tube bending structure, based on this structure, with MidasT. 8 software, simulate and an- alyze steel tube cold roll forming, set up springback data model with three input parameters based on neural network, select 1 000 set of simulated data as training data to train the neural network ,500 sets of analog data as the test data to test the network, pre- dicted results and the sample results are compared and analyzed, the results showed that nerve network prediction model meet the error requirement, it can be used to predict springback of large diameter thick wall tubes after roll forming. This research provides a theoretical basis for the development of the adaptive the CNC tube bending with springback compensation performance.